University of Groningen
Arnoldussen, S. & K.M. de Vries 2017, A plan in place? Celtic field habitation at Westeinde
(prov. Drenthe, The Netherlands)
Arnoldussen, Stijn; de Vries, Karen
Published in:
Lunula Archaeologia protohistorica
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2017
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Arnoldussen, S., & de Vries, K. (2017). Arnoldussen, S. & K.M. de Vries 2017, A plan in place? Celtic field habitation at Westeinde (prov. Drenthe, The Netherlands). Lunula Archaeologia protohistorica, 25, 79-89.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
200 m 0 220000 220400 53 56 00 53 64 00 220100 220150 53 56 00 53 56 50 20 m 0 A B C C well s24.6 pit s24.5 granary 1 granary 2 house reconstructed Celtic field banks
Celtic field banks
barrows
excavation trenches
reconstructed Celtic field banks
Celtic field banks
excavation trenches Westeinde -
NoormansveldWesteinde - Noormansveld
A plan in place? Celtic field habitation at Westeinde
(prov. Drenthe, The Netherlands)
1. Introduction
As part of a long-term research project on the use-histories of Celtic fields (Dutch: raatakkers) in the Netherlands, the site of Westeinde - Noormansveld has been investigated between 2014 and 2016. Since the 1950’s, the site has been known for the presence of a group of barrows (suspected to be Iron Age cinerary mounds, but none had been excavated). In 1999, a restoration project aimed at clearing the barrows from trees and restoring their original contours (many had been dis-turbed in their centers) discovered and mapped pronounced banks in the vicinity of this barrow cluster.Newly available, high-resolution LIDAR data indicates that these banks were actually part of a much larger Celtic field complex of 34 hec-tares (fig. 1, B).
to small scale test-trenching as part of the Groningen Uni-versity’s archaeological training dig in 2014 and 2015. In 2016, the project partners agreed to use mechanical stripping in test-trenches of a plot that – based on earlier corings and test-pits - had been considerably disturbed by recent agricul-ture. The objectives of this campaign were twofold: from a heritage management perspective it was worthwhile to deter-mine if and to what extent features may have been preserved regardless of the evident levelling visible in LIDAR plots of the ploughed field. Secondly, within the context of the Celtic field research programme interest lay primarily in investiga-ting possible interrelations of habitation and agriculture, as a local amateur archaeologists (ms. S. van der Meulen) had found various later prehistoric remains from the field when it was still regularly ploughed.
1
1Groninger Instituut voor Archeologie, Poststraat 6, 9712ER, Groningen,
s.arnoldussen@rug.nl; karen.m.de.vries@rug.nl
Fig. 1. Location of Westeinde - Noormansveld in the Netherlands (A; the shaded areas are regions with documented Celtic field complexes, Westeinde is indicated with a red star), (B) the extent of the Celtic field banks(brown thick lines) and locations of the barrows (red circles) in relation to the excavation trenches (blue outlines) and (C) inset of the Westeinde house-site.
Due to a fruitful cooperation of the estate owners
(Vereni-ging Natuurmonumenten; mr. R. Popken), the municipal and
provincial authorities (ms. M. Nieuwenhuis and mr. W.A.B. van der Sanden respectively) and the Groningen Institute of Archaeology (mr. S. Arnoldussen), the Celtic fields and barrows of Westeinde - Noormansveld could be subjected
2. Celtic fields and houses: an awkward
rela-tionship?
Despite the fact that artist reconstructions often convey an unproblematic integration of domestic life and agriculture within the Celtic fields (fig. 2), we argue that this relation is in reality poorly understood. The depicted artist’s recon-structions show houses neatly slotted into su ciently large parcels, even with room to spare for the outbuildings close-by. The orientation of the farmhouses of the reconstructions
matches those of the local fi eld plots, and the Celtic fi eld plots act as de facto house-sites or farmyards for the houses con-tained within them. In general terms, this is what archaeolo-gists take to be true as well: Iron Age habitation is assumed to have been placed in, and periodically shifted within, the confi nes of the Celtic fi eld system they exploited (e.g. Ger-ritsen 2003: 172; Spek et al. 2003: 143; Harsema 2005: 548; Meylemans et al. 2015: 207).
Remarkably, even a cursory glance at the available evidence for settlement-Celtic fi eld interrelations suggest that this is not as clear-cut as archaeologists generally assume. The oldest site for which an overlapping of house-sites and Cel-tic fi eld plots was uncovered, is the site of Wekerom - De
e h t n i h c s r u B . C . F y b d e t a v a c x e s a w e ti s s i h T . g n o r p s fj i V
war years, yet never published by him (Van Klaveren 1986; Arnoldussen & Scheele 2014: 14). Bursch’ documentation suggests that the morphology and thus the true nature of the site was not yet understood, as he depicts the Celtic fi eld’s fi eld plots as oval embanked plots. Later analysis of aerial photographs and LIDAR data have confi rmed that here too a more or less rectangular system of embanked fi elds must have been in place (Brongers 1976: 147; Oude Rengerink 2004: 24-26). Unfortunately, it is diffi cult to re-align the 1940’s excavations with the pattern of Celtic fi eld banks as reconstructable from the LIDAR data and aerial photographs (Arnoldussen & Scheele 2014: 15). Moreover, no
strati-graphic relations were documented for the houses and banks, as a result we can only note the peculiar overlap of house plans with reconstructed Celtic fi eld banks (fi g. 3, A). Ver-wers’ interpretation (1972: 146-147) that a Celtic fi eld was constructed on top of a previous settlement site, cannot be checked. What is noteworthy, though, is that for the types of houses uncovered at Wekerom (Type Wekerom or Dalen; Ar-noldussen & Scheele 2014: 19; Waterbolk 2009: 64), radio-carbon dates spanning the period 400 BC – 0 AD have been obtained (Waterbolk 2009: 64; Schabbink 2013: 24 note 1). For the banks at Wekerom, the incorporated ceramics hint at construction around the Late Bronze Age or Early Iron Age (Arnoldussen & Scheele 2014: 80), but the OSL dates obtained for the Wekerom banks suggest that bank construc-tion may have started at the end of the Middle Bronze Age and was well underway by the Early Iron Age (Wallinga & Versendaal 2013a: 6). This indicates that at Wekerom, ha-bitation is much more likely to have followed than preceded Celtic fi eld usage.
At the site of Peelo - Kleuvenveld, three Iron Age house plans and several granaries were excavated. The northwesternmost house was radiocarbon-dated to the period of c. 710-450 cal. BC (Kooi 1997: 422). The two other houses could only be dated based on typological grounds: we feel that the Early Iron Age date initially (Kooi & De Langen 1987: 58(158) proposed for both houses fi ts the evidence best (cf.
Water-Fig. 2. Cut-outs of artist reconstructions of Celtic fields, showcasing the suspected interrelations of houses and fields (clockwise from top-left; © Gemeente Ede, © Drents Museum, © Provincie Drenthe, © S. Drost).
174000 45 77 50 235250 235500 55 92 50 55 95 00 228500 229000 54 70 00 100 m 0 0 100 m 100 m 0 Wekerom - Lunteren A Peelo - Kleuvenveld Hijken - Hijkerveld C B H18
Fig. 3. Overview of excavated Celtic fields with Iron Age habitation (A: Wekerom - Lunteren: after Van Klaveren 1986; Arnoldussen & Scheele 2014: 15 fig. 8), B: Peelo - Kleuvenveld (after: Kooi & de Langen 1987; Kooi 1997), C: Hijken - Hijkerveld (after: Harsema 1974; 1991: 23 fig. 2; Arnoldussen & De Vries 2014: 101 fig. 12), all to the same scale). White areas and outlines represent the excavated areas, the locations
of reconstructed and observed Celtic field banks are depicted in halftone brown. Iron Age houses and outbuildings are depicted in red and the black polylines represent fence lines (note that for Peelo and Hijken barrows also present have been omitted from the plans).
bolk 2009: 49; 53 fi g. 31; 54; 56 fi g. 32, contra Kooi (1997: 434) who postulated a Late Iron Age date for the northern-most house, yet see Kooi 1997: 459). Two of the Kleuvenveld houses are situated amidst Celtic fi eld banks, which were – in absence of dating evidence – assigned a generic Iron Age date (Kooi & De Langen 1987: 55(155); 60(160); Kooi 1997: 418). Kooi and De Langen (1987: 61(161)) noted that the ori-entation of some of the easternmost granaries matched those of the Celtic fi eld banks nearby, implying a (temporal) asso-ciation, but they postulated that the house-sites could either have been spared during later Celtic fi eld development or al-ternatively have been constructed within a pre-existing Celtic fi eld (loc. cit.). In the later publication by Kooi (1997: 463) a functional relation (i.e. implying contemporaneity) for the Early Iron Age house-sites and fi elds is implicitly suggest-ed (Kooi 1997: 462 fi g. 39; 463). Either way, the placement of two of the Peelo - Kleuvenveld house-plans does deviate from the reconstructions presented in fi g. 2: the houses are
placed in small plots and partly overlap with the locations of the banks.
A similar observation – of seemingly illogical placement of Iron Age house plans with regards to the bank locations – could be documented at Hijken - Hijkerveld (fi g. 3, C: Har-sema 1991; Arnoldussen & De Vries 2014). Save for one house (H18; datable to the Middle to Late Iron Age: Ar-noldussen & De Vries 2014: 99; ArAr-noldussen & Brusgaard 2015), all Iron Age houses could be dated to the Early to Middle Iron Age (based on pottery and radiocarbon dates (Arnoldussen & De Vries 2014: 93). For Hijken, an impor-tant observation is that between the Iron Age houses, net -a v n i h c i h w , d e r e v o c s i d e r e w s e c n e f k r o w e lt t a w f o s k r o w
rious locations overlapped with Celtic fi elds banks as visible through soil discolouration at the time of excavation (op.cit: 101 fi g. 12). Possibly, the fence lines formed a blueprint for the systems of banks (Harsema 1980: 20-21; 1987: 39), but
Stijn A & Karen M. V
the fi eldwork notes suggest some fence lines topped banks as well (Arnoldussen & De Vries 2014: 100). Unfortunately no stratigraphic information or dating for the Hijken Celtic fi eld (banks) is available, and the georeferencing of the le-gacy data is diffi cult (loc. cit.). Regardless of the exact ages of the banks, again the positioning of the houses in rela-tion to the banks is peculiar: four house-plans overlap with documented banks and three more houses overlap with the locations of banks based on aerial photographs. Moreover, only few (three?) houses appear to be conveniently situated amidst fi elds, yet locations overlapping with banks appear to be more common here.
The three examples discussed above show clear similarities. First, absolute dates are scarce for both houses and Celtic fi eld banks alike. The latter could be dated by OSL at Weke-rom, and suggested that habitation here was much younger than the start of Celtic fi eld bank development. For Peelo and Hijken, contemporaneity was suggested, irrespectively of the awkward placement of houses in relation the Celtic fi eld banks. There is evidence to suggest that the dates obtained for bank deve lopment at Wekerom are not unique: the Celtic fi eld of Zeijen – Noordse veld was also subjected to OSL dating which indicated bank construction after the end of the Middle Bronze Age-B and continuing throughout the entire Iron Age (Wal linga & Versendaal 2013b: 6). At Someren - De Hoen-derboom, AMS and OSL dates confi rm bank development between the end of the Middle Bronze Age-B and the Late Roman period (not yet published). If we assume that the few obtained dates are representative for the development of Cel-tic fi eld banks in general, the above examples suggest that the habitation traces date to the developed or fi nal use-phases of the Celtic fi eld system. In other words: if Celtic fi elds banks came into being from the Middle Bronze Age to Late Bronze Age transition (i.e. c. 1200-1000 cal. BC) onwards, where are the houses dating to those phases? Middle- or Late Bronze Age house-plans have not been found in any Celtic fi eld ex-cavation, other than at Hijken (where their orientation may have inspired la ter Celtic fi eld orientation; Arnoldussen & De Vries 2014: 10). When datable, the house-sites invariably prove to date to va rious phases of the Iron Age, suggesting that these houses were constructed amidst pre-existing fi elds and banks. The fact that the orientation of the farmhouses in the three examples quoted (fi g. 3) conforms to those of the Celtic fi eld in a general sense, could suggest that the layout of the agricultural landscape was visible, and mattered to, the Iron Age farmers who constructed their houses there. Yet, contrary to what the artists’ reconstructions convey, no deli-berate ("logical") placement of houses in the middle of fi elds is discernible. Rather, banks seem to be as suitable – if not more – than fi eld plots for Iron Age housing.
The above discussion has highlighted that (a) excavation and study of Celtic fi eld systems in general and (b) specifi cally the interrelations with habitation within are much needed. Therefore, in the summer of 2015 a campaign of test-trenches was undertaken at Westeinde to investigate the presence and age of any habitation possibly preserved. In trench 24 (fi g. 1, C), a later prehistoric house-site was discovered, of which the dating is discussed below.
3. Celtic fi eld habitation at Westeinde
Although models suggest that Celtic fi elds systems were used for both agriculture and habitation, the actual discov-ery of a house plan situated within visible traces of a Celtic fi eld complex is not common (e.g. De Wit et al. 2003: 141; Schrijer & De Neef 2008, 78). At the most southern end of trench 24, the remnants of a small house were uncovered. The feature density was low, which allowed for good recognition and registration of the features. The uncovered house plan measured roughly 12.5 by 7 meters (fi g. 4). The outer posts (rafter supports placed outside the wall proper) were evenly spaced with at least one recognisable entrance on the northern long side of the house. The entrance in the southern wall was less clearly recognisable, because of possible renovations of this entrance. In contrast to the easy recognition of the house plan as a whole, the interpretation of the inner structure was less evident. It seems that the inner structure combined a two-aisled and three-two-aisled arrangement of roof-bearing supports. The roof-bearing posts in the smaller part of the house dis-played a T-shaped arrangement, whereas the larger (byre?) part of the house was both two- and three aisled.
Although the house plan was clearly recognisable as such, it was built in an uncommon way. That is to say, the house plan itself did not fi t well within the present house typologies of the northern Netherlands (Huijts 1992; Waterbolk 2009). Even though some characteristics of the Westeinde house are also seen with traditional types, its attribution to a specifi c type remained problematic. The evenly spaced posts placed outside the wall (supports for beams on which the rafters rest-ed), for example, are a feature of the Een-type (traditionally dated into the Early Iron Age: Waterbolk 2009: 54, 56, fi g. 32), but the house plan lacks the systematic three-aisled, in-ner confi guration of roof-bearing posts that Een-type houses usually have. In the same way, the combination of a two- and three-aisled plan is characteristic of the Dalen- and Di phoorn-types (traditionally dated into the Middle to Late Iron Age: Waterbolk 2009: 64, fi g.40-41), but the Westeinde house plan does not have the associated double wall posts typical for these two types. This is all the more remarkable, as the even-ly spaced wall posts indicate a well thought out plan and not an ad hoc solution.
Based on the ceramics recovered from the postholes of the house plan, a terminus ad or post quem in the Late Bronze Age or Early Iron Age seems likely. As all fragments from the postholes were small (10 gram or less), the possibility remains that the ceramics were mere inclusions of material already present at the site when the house was built. In one of the central posts (S84; fi nd no. 1265), two potsherds were found with a sharp angular pot profi le. Just above the off set a horizontal row of fi ngertip or nail imprints was applied (fi g. 5: v1265). This type of decoration is often found in Urnfi eld contexts (e.g. Kooi 1979: 34, fi g. 24.113, 60, fi g. 51.109). An urn from Oosterwolde (province of Drenthe) with simi-lar shape and decoration was 14C-dated to 1111-834 cal. BC
(GrN-10441: 2805 ± 55 BP; Lanting & Van der Plicht 2003: 217). In the unpublished settlement site of Welsum, a pit was found with sherds of a biconical vessel with similar
deco-Fig. 4. Feature types (top) and depths (bottom) for the remains of the Westeinde house plan (for the location of the house see deco-Fig. 1). Depth of features 0 40 cm 220090 220100 53 55 70 53 55 80 5 m 0 220090 220100 53 55 70 53 55 80 5 m 0 Features Houseplan Pit or posthole from other phase Pit or posthole, possibly natural Disturbed
S32
Stijn A & Karen M. V
rations, that was dated to between 1277-918 cal. BC (GrN-7600: 2905 ± 65 BP; Lanting & Van der Plicht 2003: 204). This combination of vessel type and decoration is thus found from the Late Bronze Age onwards. Other ceramics associ-ated with the house could not be identifi ed as specifi c vessel shapes. The properties of the fragments (granite temper, a smoothed or polished surface and thoroughly fi red) suggest a dating somewhere in the period Late Bronze Age to Mid-dle Iron Age (Taayke’s (1995: 71-72) G0/G1-types). Charred material from the central posthole was carbon-dated (S 32; fi nd no. 1264). Unfortunately, the samples must have con-tained older material as the 14C-date was between 3090 and
2910 cal. BC (Beta-45440: 4380 ± 30 BP).
As mentioned above, the Westeinde house plan cannot easily be placed within the traditional typological framework. The parallels found either came from the southern parts of Dren-the or – even furDren-ther south – from Dren-the province of Overijssel. At the excavation of Zwinderen - Kleine Esch (Van der Velde
et al. 1999), a house was excavated which was comparable in
size and construction to the Westeinde house, as it had both evenly spaced wall posts and similar inner structure (structure 5: Van der Velde et al. 1999: 71-73). Waterbolk (2009: 56, fi g. 32e) interpreted it as an Een-type, but attribution is not com-pletely convincing. The house itself did not yield any datable fi nds, but in the same trenches fi nds from both the Middle Iron Age and 4th and 5th century AD were found (Van der
Vel-de et al. 1999: 71, 76). From Raalte-Jonge Raan (pro vince of Overijssel) another similar structure was found (house 3: Groenewoudt et al. 1998: 36-37). Its wall construction is simi lar to that of the Westeinde house, although the house plan from Raalte - Jonge Raan is slightly larger (16 meters prior to extension). Again, the confi guration of the roof-bear-ing posts represents a combination of both two- and three-aisled building techniques. Ceramics found in a pit inside the Raalte - Jonge Raan house suggest a dating at the start of 1st
century AD. This is confi rmed by 14C-dates of charred cereals
from the same pit, which date between 94 cal. BC and 118 cal. AD (1990 ± 40 BP: GrA-9449; Groenewoudt et al. 1998: 25, fi g. 3.3). Based on these examples, it becomes evident that these small houses are not typical for just one period or sub-period. Rather, it seems that there was a continuation of a more generic Iron Age building tradition that co-existed with other – better recognisable – house types.
The Westeinde house plan is not unique in its deviation from the traditional house typology, as house plans at other sites also do not always fi t neatly. For example, at Hijken - Hijk-erveld a number of Early to Middle Iron Age houses showed general traits of an Iron Age building tradition without a pre-cise correspondence to a type (Arnoldussen & De Vries 2014: 92-95). Similarly, the northernmost Early Iron Age sites of Angelslo - Emmerhout yielded two similar house-plans that could not easily be fi tted within the traditional typologies (Kooi 2008: 335 fi g. 4; 361, cf. Huijts 1992: 62-65). Another example is Structure 6 at Borger - Daalkampen Klokbeker (Van der Meij 2010: 25-27), which also showed traits of two diff erent types. This presence of characteristics of diff erent house types made the excavator even raise the question what constructive elements should be decisive in
the attribution to a type (Van der Meij 2010: 26). The solu-tion to this problem lies in the recognisolu-tion that houses are composites of constructive elements (e.g. wall, roof-carry-ing construction, entran ces; cf. Arnoldussen 2008: 192-198). Constructional elements may co-occur in standard ways that makes a "type" reco gnisable, but sometimes elements are found in more unique confi gurations as seems to be the case at Westeinde.
Close to the Westeinde house, two granaries were found. The granary closest to the house (granary 1, fi g. 1C) had a similar orientation as the house itself. The carbon-dates of charred material from one of the postholes (S19, fi nd no. 1136), ho-wever, dated to between 1745 and 1620 cal. BC (Beta-45441: 3390 ± 30 BP). Considering the conformity in orientation of granary 1 and the house, the dated material represents unin-tentionally incorporated evidence of Middle Bronze Age ac-tivities in the area and has little relevance to the structure into which it became incorporated.
During the campaign of 2016, a total of 1456.6 grams of ce-ramics were recovered from 49 features. The average weight of the ceramics per feature was less than 30 grams. One pit (S5, fi nd nos. 826, 1121, 1188, 1191 and 1281) stood out from the other features as it contained nearly half of all ceramics found (717 grams). The pit was round and shallow (diameter 80 cm, maximum depth 15 cm) with a fl at base. The fi ll of the pit was homogeneous and with a high amount of char-coal. The original function of the pit remains unclear, but its fi ll comprised several layers of dark and light grey material, which may have been charcoal and ash, suggesting the incor-poration of remains from a hearth or cooking pit. The lack of burned sand underneath the pit at least indicates that the pit was not used for cooking with large open fi res.
After the primary use of the pit, the content must have at least partially been removed. A band of dark and more humic material can be seen several centimetres inwards from the edges of the original pit. Remarkably, these humic layers are also visible at the vertical edges of the pit. This suggests that material was placed against the walls of the pit, which was then burned (cf. the sanitation of silo pits through burning; Arnoldussen 2008: 263 note 296). From three diff erent seg-ments of the pit, samples were taken and sieved. Except for one charred grain of Emmer wheat, no botanical remains of plants – other than small fragments of charcoal – were found in the fi ll of S5.
The fi nds from S5 consisted of 55 potsherds, with a total weight of 717 grams. Apart from three fragments that will be discussed below, the remaining fragments seem to have belonged to the same (type of) vessel. A large rim fragment could be reconstructed, consisting of 15 sherds. The vessel has a small, thick rim (fi g. 5, top). The upper part of the body of the vessel is smoothed, almost polished, while the lower part of the body is smitten (besmeten). The tempering of the pot – fi ne stone grit temper with only minor inclusions of an organic temper – and polished surface at the neck of the vessel suggest a Middle Iron Age date (Taayke’s G3-type; Taayke 1995: 16). The shape of the vessel is less conclusive,
because it has characteristics of both the G3-type (Middle to Late Iron Age: Taayke 1996: 182, fi g. 10.d) and the G5-types (Early Roman period: loc.cit.). The Westeinde vessel from S5 has the simple, thick rim of the G3 type, but its dimensions (diameter at the rim: 27 cm) are uncommonly large (pers. comm. E. Taayke, December 2016). While shape and size are fi tting for the younger G5-type, the vessel lacks the protru-ding and faceted rim of the G5-type.
The other three fragments are rim fragments belonging to an open shaped vessel or bowl. The bake of the rim fragments comparable to that of the vessel discussed earlier: a fi ne stone grit temper, with only minor incorporations of an organic temper. The top of the rim is decorated with fi ngertip imprints (fi g. 5, lower left). Bowls decorated with fi ngertip impres-sions are not commonly found, although some examples are known. For example, fragments of similar type of bowls were found at Zeijen - Witteveen (Waterbolk 1977: 17 (189), fi g. 8.23-25) and further north in the terp region, from the earliest phases of Ezinge (Nieuwhof 2014: 38, fi g. 5.1602-14). This type of decorated bowls is also known from Oss-Ussen, in the southern parts of the Netherlands (Van den Broeke 2012, 46-50, esp. fi g. 3.5.4-5 and fi g. 3.6.7, 3.6.13). The bowl frag-ments from Zeijen and Ezinge date into the Middle Iron Age. In similar vein, in Oss - Ussen the bowl type is common-ly found up to the second half of the Middle Iron Age, after which there is a steep decrease in numbers. The type does show a small revival in the Roman period, although without
0 5 cm Wp24 S5 v1191 Wp24 S82 v1265 Wp24 S5 v1191
Fig. 5 Part of the ceramics recovered from the pit (S5: v1191) and from a roof-bearing post of the house (S84: v1265), all to the same scale.
the fi ngertip imprints on the rim (Van den Broeke 2012, 46 fi g. 3.5 nos. 4-5). A carbonised Triticum grain was submitted for AMS dating, but unfortunately did not yield suffi cient car-bon to allow dating. A provisional dating into the Middle Iron Age is suggested here.
All but six ceramics sherds clustered in the northernmost part of the pit (fi g. 6, top). The majority of the fragments were found in the upper 10 centimetres of the fi ll. Ten sherds were found just on top of the charcoal rich layer. What stands out is the diff erent degrees of weathering of the sherds that cannot be explained by usage (i.e. diff erential use-wear) of the pot. The lower sherds of the refi tted fragment show quite extensive wear, as the originally present smitten surface was completely erased from the vessel (see fi g. 6, inset). Adja-cent sherds show no traces of such weathering. It is also not likely that these diff erences were caused by – diff erent – geo-chemical processes because all fragments were found in the same fi ll of the pit. This is important, as it suggests that indi-vidual sherds may have had diff erent use lives prior to them being placed in the pit. Diff erences also exist between what fragments were found where. All small sherds were found in the upper fi ll, whereas the relative larger fragments were all found at the bottom. The fact that they were not found in articulation argues against the rim fragment being broken under pressure of the fi ll. The sherds ended up being deposi-ted in the pit, although restricdeposi-ted to one specifi c locale within the pit.
Fig. 6. Pit S5 with the ten pottery fragments found just on top of the charcoal rich layer (top). Refi tted fragment of a large ferences in weathering on adjacent sherds (bottom right). Numbers in top and bottom photo are corresponding and referring to the same sherds.
4. A plan in place: the implication of the
West-einde house-site
The excavation of Westeinde can be added to the list of set-tlement sites that do not neatly fi t with the present models of Iron Age habitation. This ill-fi t holds for the present un-derstanding of both Iron Age house building traditions and the ways houses are placed within Celtic fi eld systems. The Westeinde house was probably built during or after the Late Bronze Age. Although it cannot be dated more precisely at the moment, it is dissimilar to any later prehistoric house plan that has been uncovered in the region thus far. Even though
the house is unique in its construction, its constructional ele-ments are also not completely alien to the region. Di? erent characteristics of the house can be recognised across several other later prehistoric structures. In this way, the Westeinde house-plan shows once more that it is impossible – and possi-bly even unhelpful – to describe all house-plans according to traditional typologies.
The absolute dates obtained for the Westeinde house and out-buildings could not confi rm contemporaneity between the house and the nearby outbuildings. Two scenarios remain possible: in the fi rst scenario a small farmstead with house and G3-vessel (bottom left) and dif
(1)
(2)
(3)
(4)
(2)
(4)
(1)
(3)
two outbuildings was situated amidst the Celtic fi eld. In the second scenario the outbuildings and farm did not belong to the same phase. In both cases, this means that not all domestic activities were restricted to one and the same plot. Even if the house and granaries were contemporary, a nearby uninha-bited fi eld plot at some distance of the house comprised a pit containing pottery and a well (fi g. 1C). Even though at this moment their temporal association with the house could not be confi rmed, the pit and well are found outside the confi nes of a "single-plot" farmstead within the Celtic fi eld. Evidently, the Westeinde excavation results contradict the existing im-age of a domestic unit within a otherwise fully agricultural system, as portrayed in fi gure 2. Activities of a domestic na-ture (cooking, fetching water) may have been spread across various Celtic fi eld plots and were not restricted to the sin-gle plot containing the house proper. This also challenges the existing image of what exactly comprises a farmstead. The importance of easy access to outbuildings may have been the reason these were located close to the house, but this may have been diff erent for wells and cooking pits. In addition to this, the farmstead may not have been a clearly defi ned and demarcated unit, but rather a more loosely structured zone around the house within which all necessary elements (shel-ter, drinking wa(shel-ter, etc.) for daily life could be found. The Westeinde site is no exception in this, as other Celtic fi elds show similarly dispersed domestic elements (e.g. Sellingen: Van Giff en 1939 and Gees: Waterbolk 1989: 290, fi g. 2). The similarity in orientation of the house-plan and one of its outbuildings (granary 1; fi g. 1, C) to the general orientation of the system of banks supports a certain degree of contempora-neity, but still allows a sequence in which later Iron Age ha-bitation was slotted into a cultural landscape mark-up of con-siderably older age – which nonetheless still provided visual clues as to its dominant axis of orientation. This conformity of house-orientation and bank-orientations is also visi ble at Wekerom, Peelo and Hijken (fi g. 3), and several other sites (e.g. Sellingen, Zeijen - Noordseveld; Van Giff en 1939: fi g. 3a; Waterbolk 1977). Seemingly paradoxically, whereas the
orientation of the banks is respected and mirrored by the
houses, their location is not. At Westeinde, Sellingen, Hijken, Peelo and Wekerom, some of the house-plans overlap with the location of banks, but due to absence of stratigraphic informa-tion their phasing cannot be determined. In a general sense, with OSL and radiocarbon dates suggesting a start of the bank construction from the end of the Middle Bronze Age-B onward (Wallinga & Versendaal 2013a-b), it seems plausible that the houses were built after bank construction – as several Early (Hijken, Peelo) to Middle/Late (Vaassen, Wekerom) dates for the houses have been proposed. Clearly, micro-topographic diff erences in height did not render such locations unfi t for habitation, and one may wonder whether bank locations were favoured as house-sites (albeit that the overlap of house-plans and banks could have been better had prehistoric communities strived to place all houses on bank apexes).
Estimating the number of house-plans potentially present within the given Celtic fi eld complex is diffi cult. One could argue that the location of the test-trenches at Westeinde un-covered only 7 fi elds suffi ciently extensively to decisively
argue whether or not they contained houses. At Vaassen and Peelo, c. 8 and 5 fi eld plots were suffi ciently investigated with test-trenches to determine the presence or absence of hou
-ses (found in 1 and 3 fi eld plots respectively; Brongers 1976: 44-45 fi g. 4; Kooi & De Langen 1987: 58(158)). Similarly, the excavations at Wekerom - Lunteren uncovered 5 houses within the 21 fi eld plots investigated (Arnoldussen & Scheele 2014: 15 fi g. 8). At Hijken, no less than 8 house-sites were found within 19 investigated fi eld plots (Arnoldussen & De Vries 2015: 101 fi g. 12). At Sellingen, Van Giff en uncovered a house-plan underneath the bank subdividing two fi eld plots (Van Giff en 1939: 90). These few cases suggest that the ratio of fi eld plots with (Iron Age) habitation versus total of fi eld plots ranges between 0.11 and 0.5, with a mean of 0.24 for those site with more than 5 fi eld plots investigated (st.dev. 0.12). In other words, for every inhabited plot there were four plots not inhabited. Extrapolating this ratio of 0.24 for the Westeinde Celtic fi eld complex, suggests that the over 200 fi eld plots of Westeinde (fi g. 1, B) may harbour 40 to 50 of such house-sites. Taking into account the estimates for re-quired areas of arable for 5-8 person Iron Age households practising mixed agriculture (1.73 - 2.4 ha; Fokkens 1998: 144 Tab. 27; Woltering 2000: 343 Table 18), the modelled "carrying capacity" of the 34 ha Westeinde complex is 12-19 house-sites, which means that several Iron Age use-phases with contemporary house-sites are to be expected. Unfortu-nately, due to the small number of fi elds presently investiga-ted at Westeinde such extrapolations must presently remain very tentatively.
References
ARNOLDUSSEN S., 2008. A Living Landscape: Bronze Age
settlement sites in the Dutch river area (2000-800 BC).
Lei-den (PhD thesis): Sidestone press.
ARNOLDUSSEN S. & BRUSGAARD N.Ø., 2015. Produc-tion in deposiProduc-tion: structured deposiProduc-tion of Iron Age iron-working elements (The Netherlands). Lunula Archaeologia
protohistorica, XXIII, pp. 115-124.
ARNOLDUSSEN S. & SCHEELE E.E., 2014. De Celtic
Fields van Wekerom: kleinschalige opgravingen van wal-len en velden van een laat-prehistorisch raatakkersysteem
(Grondsporen 18). Groningen: Groningen Institute of Ar-chaeology (GIA).
ARNOLDUSSEN S. & VRIES K.M. DE, 2014. Of farms and fi elds. The Bronze Age and Iron Age settlement and Celtic fi eld at Hijken - Hijkerveld. Palaeohistoria, 55/56
(2013/2014), pp. 85-104.
BROEKE P. W. VAN DEN, 2012. Het handgevormde
aardewerk uit de IJzertijd en Romeinse Tijd van Oss-Ussen. Studies naar typochronologie, technologie en herkomst.
Lei-den (PhD thesis): Sidestone press.
BRONGERS J.A., 1976. Air photography and Celtic fi eld
research in the Netherlands (Nederlandse Oudheden 6).
Stijn A & Karen M. V
FOKKENS H., 1998. Drowned Landscape. The Occupation
of the Western Part of the Frisian-Drenthian Plateau, 4400 BC - AD 500. Assen: Van Gorcum.
GERRITSEN F., 2003. Local identities. Landscape and
com-munity in the late prehistoric Meuse-Demer-Scheldt region
(Amsterdam Archaeological Studies 9). Amsterdam: Amster-dam University Press.
GIFFEN A.E. VAN, 1939. De zgn. heidensche legerplaats te Zuidveld bij Sellingen, gem. Vlachtwedde. Verslag
Gro-ninger Museum, 1939, pp. 86-93.
GROENEWOUDT B., SPEK TH., VELDE H. M. VAN DER, AMEN I. VAN, DEEBEN J.H.C. & SMEERDIJK D.G. VAN, 1998. Raalte-Jonge Raan: de geschiedenis van een Sallandse
bouwlandkamp (Rapportage Archeologische
Monumenten-zorg 58). Amersfoort: Rijksdienst voor het Oudheidkundig Bodemonderzoek.
HARSEMA O.H., 1974. Archeologisch onderzoek op het Hijkerveld, gemeente Beilen. Voorlopig bericht over de op-gravingen 1969 en 1970. Nieuwe Drentse Volksalmanak, 91, pp. 161-168.
HARSEMA O.H., 1980. Drents boerenleven van de
brons-tijd tot de middeleeuwen (Museumfonds Publicatie 6). Assen:
Drents Museum.
HARSEMA O. H., 1987. De archeologische betekenis van
het Hijker Noorderveld en zijn omgeving. Groningen: BAI
(unpublished report).
HARSEMA O.H., 1991. De bronstijd-bewoning op het Hij-kerveld bij Hijken. In: H. FOKKENS, and N. ROYMANS (eds.). Nederzettingen uit de bronstijd en de vroege ijzertijd
in de Lage Landen. (Nederlandse Archeologische Rapporten.
Amersfoort: Rijksdienst voor het Oudheidkundig Bodem-onderzoek, pp. 21-29.
HARSEMA O.H., 2005. Farms amongst Celtic fi elds. Set-tlements on the northern sands. In: L. P. LOUWE KOOIJ-MANS, P. W. V. D. BROEKE, H. FOKKENS, and A. L. V. GIJN (eds.). The prehistory of the Netherlands (II). Amster-dam: Bert Bakker, pp. 543-555.
HUIJTS C.S.T.J., 1992. De voor-historische boerderijbouw
in Drenthe; reconstructiemodellen van 1300 vóór tot 1300 na Chr. Arnhem: Stichting Historisch Boerderij Onderzoek.
KOOI P.B., 1979. Pre-Roman urnfi elds in the north of the
Netherlands. Groningen (PhD thesis): Wolters-Noordhof.
KOOI P.B., 1997. Het project Peelo. het onderzoek van het Kleuvenveld (1983, 1984), het burchtterrein (1980) en het Nij land (1980), met enige kanttekeningen bij de resultaten van het project. Palaeohistoria, 37/38 (1995/1996), pp. 417-479. KOOI P.B., 2008. Nederzettingen uit de bronstijd en ijzertijd in Angelslo-Emmerhout. Palaeohistoria, 49/50, pp. 327-373.
KOOI P.B. & LANGEN G.J. DE, 1987. Bewoning in de vroege ijzertijd op het Kleuvenveld te Peelo, gem. Assen.
Nieuwe Drentse Volksalmanak, 104, pp. 151-165.
LANTING J.N. & VAN DER PLICHT J., 2003. De 14
C-chro-nologie van de Nederlandse pre- en protohistorie IV: de brons-tijd en vroege ijzerbrons-tijd. Palaeohistoria, 43/44, pp. 177-261. LANTING J.N. & VAN DER PLICHT J., 2006. De 14
C-chro-nologie van de Nederlandse pre- en protohistorie V: de mid-den- en late ijzertijd. Palaeohistoria, 45/46, pp. 241-427. MEIJ L. VAN DER, 2010. Project bestemmingsplan
Daalkampen II te Borger. Locatie: Klokbeker (ADC Rapport 2065). Amersfoort: ADC archeoprojecten.
MEYLEMANS E., CREEMERS G., DE BIE M. & PAESEN J., 2015. Revealing Extensive Protohistoric Field Systems through High Resolution LIDAR Data in the Northern Part of Belgium. Archäologisches Korrespondenzblatt, 45, pp. 197-213.
NIEUWHOF A. (eds.), 2014. En dan in hun geheel. De vond-sten uit de opgravingen in de wierde Ezinge. (Jaarverslagen van de Vereniging voor Terpenonderzoek 96). Groningen: Vereniging voor Terpenonderzoek.
OUDE RENGERINK J.A.M., 2004. Beheersgebied
Weker-omse Zand: een inventariserend archeologisch onderzoek: kartering en waardering Celtic Fields (RAAP Rapport 1019).
Weesp: RAAP Archeologisch Adviesbureau.
SCHABBINK M., 2013. Kleine boeren uit de IJzertijd en
de 18e/19e eeuw op ‘t Spikkert te Weerselo: Gemeente Din-kelland : archeologisch onderzoek: een opgraving. (RAAP
Rapport 2508). Weesp: RAAP Archeologisch Adviesbureau. SCHRIJER E. & DE NEEF W., 2008. Fluitenberg,
begraaf-plaats Zevenbergen. Gemeente Hoogeveen (Dr.). Deel 2: Op-graving en archeologische begeleiding (Steekproefrapport 2007-02/02). Zuidhorn: De Steekproef bv.
SPEK TH., GROENMAN-VAN WAATERINGE W., KOOIS TRA M. & BAKKER L., 2003. Formation and land-use history of Celtic fi elds in north-west Europe - An inter-disciplinary case study at Zeijen, The Netherlands. European
Journal of Archaeology, 6 (2), pp. 141-173.
TAAYKE E., 1995. Die einheimische Keramik der nördlichen Niederlande, 600 v.Chr. bis 300 n.Chr., Teil II: Nord-Drenthe.
Berichten van de Rijksdienst voor Oudheidkundig Bodem-onderzoek, 41, pp. 9-102.
TAAYKE E., 1996. Die einheimische Keramik der nördli-chen Niederlande, 600 v.Chr. bis 300 n.Chr., Teil V: Über-sicht. Berichten van de Rijksdienst voor Oudheidkundig
Bo-demonderzoek, 42, pp. 163-208.
VELDE H.M. VAN DER, HAASTER H. VAN, SPEK TH. & TAAYKE E., 1999. Archeologisch onderzoek langs de
snel-weg. Opgraving in het kader van de aanleg van de Rijksweg 37: de essen van Wachtum en Zwinderen (ADC Rapport 11).
Bunschoten: Archeologisch Diensten Centrum.
VERWERS G.J., 1972. Das Kamps Veld in Haps in
Neoli-thikum, Bronzezeit und Eisenzeit (Analecta Praehistorica
Lei-densia 5). Leiden (proefschrift).
WALLINGA J. & VERSENDAAL A.J., 2013a. NCL-7213
Quartz Luminescence Dating Report Wekerom - Lunteren.
Wageningen: Netherlands Centre for Luminescence Dating. WALLINGA J. & VERSENDAAL A.J., 2013b. NCL-7212
Quartz Luminescence Dating Report Zeijen - Noordse veld.
Wageningen: Netherlands Centre for Luminescence Dating. WATERBOLK H.T., 1977. Opgravingen rond het Witteveen op het Noordse Veld bij Zeijen, gem. Vries (1949-1953).
Nieuwe Drentsche Volksalmanak, 94, pp. 5(177)-31(203).
WATERBOLK H.T., 1989. Siebzig Jahre archäologische Siedlungsforschung durch das Biologisch-Archäologis-che Institut der Universität Groningen. Under besonderer Berücksichtigung der Untersuchungen Albert Egges van Giff en in der Provinz Drenthe. Siedlungsforschung
Archäolo-gie-Geschichte-Geographie, 7, pp. 285-320.
WATERBOLK H.T., 2009. Getimmerd verleden. Sporen van
voor- en vroeghistorische houtbouw op de zand- en klei-gronden tussen Eems en IJssel. Utrecht: Matrijs.
WIT M.J.M. DE, 2003. Een Defi nitief Archeologisch
Onder-zoek langs de Frieslandweg in Emmen, gemeente Emmen (Dr.) (ARC Publicatie 64). Groningen: ARCbv.
WOLTERING P.J., 2000. Occupation History of Texel, IV: Middle Bronze Age - Late Iron Age. Berichten van de
Rijks-dienst voor het Oudheidkundig Bodemonderzoek, 44, pp.
